Method for constructing defect detection model and defect detection method and device

Abstract

The invention discloses a method for constructing a defect detection model, and a defect detection method and device. The method for constructing the defect detection model comprises the following steps: acquiring a target defect image of a display panel; according to the brightness data of each pixel in the target defect image, constructing a row vector sample and a column vector sample, and respectively carrying out classification labeling on the row vector sample and the column vector sample; constructing a classification model for defect detection, wherein the classification model comprises a support vector machine classification model in a row direction and a support vector machine classification model in a column direction; respectively acquiring a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction; and generating a defect detection model storing the first classification hyperplane parameter and the second classification hyperplane parameter. According to the method for constructing the defect detection model and the defect detection method and device provided by the invention, local defects in the display panel can be detected.

Description

Technical Field

[0001] The present application relates to the technical field of display panel manufacturing, and in particular to a method for constructing a defect detection model, a defect detection method and a device. Background Art

[0002] Active-matrix organic light-emitting diode (AMOLED) screens are becoming increasingly popular among consumers due to their low power consumption and good color rendering effects. However, during the manufacturing process, workers' fingerprints, dust, and disturbances in production line machines may cause specific defects in the display panel, namely, panel-inherent mura. The defect has obvious characteristics and its location is uncertain.

[0003] Defect detection in the prior art is mainly aimed at global mura compensation, and generally does not involve local mura detection. However, the problem of local defects in display panels is inevitable in production lines. Therefore, it is necessary to provide an improved technical solution to detect local defects in display panels. Summary of the invention

[0004] The purpose of the present application is to provide a method for constructing a defect detection model, a defect detection method and a device, which can detect local defects in a display panel.

[0005] To achieve the above objectives: In a first aspect, an embodiment of the present application provides a method for constructing a defect detection model, comprising: Acquire a target defect image of a display panel; Constructing row vector samples and column vector samples according to the brightness data of each pixel in the target defect image, and classifying and labeling the row vector samples and column vector samples respectively; Constructing a classification model for defect detection, wherein the classification model includes a support vector machine classification model in a row direction and a support vector machine classification model in a column direction; According to the row vector samples, the column vector samples, the classification labeling results of the row vector samples and the classification labeling results of the column vector samples, respectively obtaining a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction; A defect detection model storing the first classification hyperplane parameters and the second classification hyperplane parameters is generated.

[0006] In one embodiment, acquiring a target defect image of a display panel includes: Acquiring an original defect image of the display panel; Determine the area with defects in the original defect image as the ROI area; The image of the ROI area is preprocessed to obtain the target defect image.

[0007] In one embodiment, preprocessing the image of the ROI region includes: adjusting the color scale of the ROI region to enhance the contrast of the target defect image; The step of adjusting the color scale of the image in the ROI area includes: Define calculation parameters in the color scale adjustment, the calculation parameters include: the black field threshold Sin and the white field threshold Hin in the ROI area, the Gamma adjustment factor Mt required for adjustment, and the adjusted output black field threshold Sout and the output white field threshold Hout; The scaling factor V1 is calculated according to the calculation parameters, and the adjustment coefficient V2 after Gamma adjustment is calculated according to the scaling factor V1, wherein ; ; The Pin is the original brightness data of the pixels in the ROI area; An output mapping function is established according to the adjustment coefficient V2 to obtain the output brightness data of the pixels in the ROI area after the color scale adjustment. The output mapping function is: .

[0008] In one embodiment, constructing row vector samples and column vector samples according to the brightness data of each pixel in the target defect image includes: The brightness data of each row of pixels in the target defect image is taken as a row vector sample, and the row vector sample set X_Ri = [pi1, pi2, ..., piM], M = COL, i = 1, ..., ROW; The brightness data of each column of pixels in the target defect image is taken as a column vector sample, and the column vector sample set X_Cj = [p1j, p2j, ..., pNj], N = ROW, j = 1, ..., COL; Among them, ROW is the number of rows of pixels in the target defect image, COL is the number of columns of pixels in the target defect image, and piM and pNj are the brightness of pixels in the target defect image.

[0009] In one embodiment, the classifying and labeling the row vector samples and the column vector samples respectively includes: Determining whether the pixel is a defective pixel according to the brightness data of the pixel in the row vector sample and the column vector sample; Classifying and labeling the row vector samples and the column vector samples according to the determination result of the defective pixels; Among them, when the classification labeling result of the row vector sample Ri_y = 1, it means that the i-th row in the target defect image is a defective row, and when Ri_y = -1, it means that the i-th row is a non-defective row; when the classification standard result of the column vector sample Cj_y = 1, it means that the j-th column in the target defect image is a defective column, and when Cj_y = -1, it means that the j-th column is a non-defective column.

[0010] In one embodiment, the support vector machine classification model in the row direction is: ; The support vector machine classification model in the column direction is: ; in; , are row vector samples and column vector samples respectively, , are the normal vectors in the row direction and the normal vector in the column direction, , is the distance between the classification hyperplane and the origin.

[0011] In one embodiment, obtaining a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction includes: The mathematical models of hard-margin support vector machines in row and column directions are established respectively. The mathematical model of the hard-margin support vector machine in the row direction is: ; The mathematical model of the column-wise hard-margin support vector machine is: ; Solve the mathematical model of the hard margin support vector machine in the row direction and the column direction respectively to obtain the optimal parameters of the first classification hyperplane , and the optimal parameters of the second classification hyperplane .

[0012] In a second aspect, an embodiment of the present application provides a defect detection method, using a defect detection model constructed by the above method, the defect detection method comprising: Acquire an image to be detected of a display panel to be detected; Inputting the image to be inspected into the defect detection model to perform defect detection on the display panel to be inspected; When it is detected that the display panel to be inspected has defects, the detected defective rows and defective columns are output.

[0013] In one embodiment, the defect detection method further includes: acquiring the position of the defect area according to the output defect row and defect column.

[0014] In a third aspect, an embodiment of the present application provides a defect detection device, including a processor and a memory storing a computer program, and when the processor runs the computer program, the steps of the method described above are implemented.

[0015] The method for constructing a defect detection model provided by the embodiment of the present application is to obtain a target defect image of a display panel; construct a row vector sample and a column vector sample according to the brightness data of each pixel in the target defect image, and classify and annotate the row vector sample and the column vector sample respectively; construct a classification model for defect detection, the classification model includes a support vector machine classification model in the row direction and a support vector machine classification model in the column direction; obtain the first classification hyperplane of the support vector machine classification model in the row direction and the second classification hyperplane of the support vector machine classification model in the column direction according to the row vector sample, the column vector sample, the classification and annotation results of the row vector sample and the classification and annotation results of the column vector sample respectively; generate a defect detection model storing the first classification hyperplane parameters and the second classification hyperplane parameters. In this way, a detection model that can automatically detect defect features can be generated to classify the brightness data of pixels in the defect image and realize the detection of local defects in the display panel.

[0016] The defect detection method provided in the embodiment of the present application obtains the image to be detected of the display panel to be detected; inputs the image to be detected into the defect detection model to perform defect detection on the display panel to be detected; when the display panel to be detected is detected to have defects, the detected defect rows and defect columns are output. In this way, it is possible to quickly detect whether there is a defective area on the display panel and locate the defective area. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flowchart of a method for constructing a defect detection model provided in an embodiment of the present application.

[0019] Figure 2 A schematic diagram of the division of the ROI area provided in an embodiment of the present application.

[0020] Figure 3This is a schematic diagram of the ROI area after color scale adjustment provided in an embodiment of the present application.

[0021] Figure 4 A flowchart of a defect detection method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0023] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0024] It should be understood that, although the terms first, second, third, etc. may be used to describe various information in this article, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of this article, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein can be interpreted as "at the time of..." or "when..." or "in response to determination". Furthermore, as used in this article, the singular forms "one", "one" and "the" are intended to also include plural forms, unless there is an opposite indication in the context. It should be further understood that the terms "comprising", "including" indicate that there are described features, steps, operations, elements, components, projects, kinds, and / or groups, but do not exclude the existence, occurrence or addition of one or more other features, steps, operations, elements, components, projects, kinds, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Thus, “A, B, or C” or “A, B and / or C” means “any of the following: A; B; C; A and B; A and C; B and C; A, B, and C.” An exception to this definition will occur only when a combination of elements, functions, steps, or operations are inherently mutually exclusive in some manner.

[0025] It should be understood that, although the various steps in the flowchart in the embodiment of the present application are displayed in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and it can be performed in other orders. Moreover, at least a portion of the steps in the figure may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and their execution order is not necessarily performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0026] It should be noted that, in this article, step codes such as S101, S102, etc. are used for the purpose of expressing the corresponding content more clearly and concisely, and do not constitute a substantial limitation on the sequence. When implementing the step, those skilled in the art may execute S102 first and then S101, etc., but these should all be within the scope of protection of this application.

[0027] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0029] See also Figure 1 , is a method for constructing a defect detection model provided in an embodiment of the present application, the method specifically comprises the following steps: Step S110: acquiring a target defect image of the display panel.

[0030] Specifically, in the embodiment of the present application, the display image of the display panel with defects is obtained as training data for the defect detection model. The display panel in the embodiment of the present application may be an AMOLED display panel, and when collecting the defect image of the display panel, the display image of the display panel in a working state may be collected by a collection device equipped with a camera.

[0031] In one embodiment of the present application, obtaining a target defect image of a display panel includes: acquiring an original defect image of the display panel; determining a region with defects in the original defect image as a ROI region; and preprocessing the image of the ROI region to obtain a target defect image.

[0032] Specifically, the original defective image may be a display image of a display panel with defects directly captured by a capture device, and the display image includes a defective area. Figure 2 As shown, the defective area is usually located in the middle area of ​​the display panel. Therefore, the ROI area (Region of Interest) containing the defective pixels can be set according to the shape of the display panel. It should be noted that the shape, position and size of the ROI area can be adaptively adjusted according to the actual application scenario to reduce the amount of calculation for subsequent image processing and model analysis and save computing resources. Figure 2 The original image of the display panel taken in the image can be used to define a rectangular area in the middle of the image as the ROI area. By preprocessing the image in the ROI area, the image quality is improved and the defect features of concern are made more obvious, so as to facilitate the subsequent extraction of defect features.

[0033] In one embodiment of the present application, preprocessing the image in the ROI region includes: adjusting the color scale of the image in the ROI region to enhance the contrast of the target defect image.

[0034] The color scale adjustment of the ROI area includes the following steps: Step S111: define the calculation parameters in the color scale adjustment, where the calculation parameters include: the black field threshold Sin and the white field threshold Hin in the ROI area, the Gamma adjustment factor Mt required for adjustment, and the adjusted output black field threshold Sout and output white field threshold Hout. It should be noted that the values ​​of the above calculation parameters can be customized according to the desired adjustment effect.

[0035] Step S112: Calculate the scaling factor V1 according to the calculation parameters, and calculate the adjustment coefficient V2 after Gamma adjustment according to the scaling factor V1, where ; ; Pin is the original brightness data of the pixels in the ROI area, and the range of Pin is [0,255].

[0036] Specifically, the scaling factor V1 is a coefficient for adjusting the image contrast, and the original brightness data of the pixels in the ROI area is linearly stretched by setting the black field threshold Sin and the white field threshold Hin. In this embodiment, V1 is protected for the maximum and minimum values ​​to ensure that V1 is within the range of [0, 255]. V2 is a new coefficient after applying the Gamma adjustment factor, and the original brightness data after linear stretching is Gamma transformed by the set Gamma adjustment factor Mt to correct overexposed or underexposed areas.

[0037] Step S113: Establish an output mapping function according to the adjustment coefficient V2 to obtain the output brightness data of the pixels in the ROI area after the color scale adjustment, wherein the output mapping function is: .

[0038] In this embodiment, the output brightness data Pout is protected from the maximum and minimum values, so that the output brightness data Pout is within the range of [0, 255]. The image data of the ROI area is mapped to the original brightness data Pin, and the corresponding output brightness data Pout is calculated to adjust the color scale. The image of the ROI area after the color scale adjustment is as follows: Figure 3 As shown, it can be found that the contrast of the image is enhanced after the color level adjustment, so that the defect features in the image are more obvious and easier to identify.

[0039] Step S120: constructing row vector samples and column vector samples according to the brightness data of each pixel in the target defect image, and classifying and labeling the row vector samples and column vector samples respectively.

[0040] Specifically, according to the resolution ROW*COL of the target defect image, the number of rows and columns of pixels in the target defect image can be obtained. Each row and each column of the target defect image includes multiple pixels. In one embodiment of the present application, the brightness data of each row of pixels in the target defect image is taken as a row vector sample, and the row vector sample set X_Ri = [pi1, pi2, ..., piM], M = COL, i = 1, ..., ROW; the brightness data of each column of pixels in the target defect image is taken as a column vector sample, and the column vector sample set X_Cj = [p1j, p2j, ..., pNj], N = ROW, j = 1, ..., COL; wherein ROW is the number of rows of pixels in the target defect image, COL is the number of columns of pixels in the target defect image, and piM and pNj are the brightness of pixels in the target defect image.

[0041] In one embodiment of the present application, classifying and labeling row vector samples and column vector samples separately specifically includes: judging whether a pixel is a defective pixel based on the brightness data of the pixel in the row vector samples and the column vector samples; and classifying and labeling the row vector samples and the column vector samples based on the judgment result of the defective pixel.

[0042] In a specific embodiment, when the classification labeling result Ri_y = 1 of the row vector sample, it indicates that the i-th row in the target defect image is a defective row, and when Ri_y = -1, it indicates that the i-th row is a non-defective row; when the classification standard result Cj_y = 1 of the column vector sample, it indicates that the j-th column in the target defect image is a defective column, and when Cj_y = -1, it indicates that the j-th column is a non-defective column.

[0043] Step S130: constructing a classification model for defect detection, the classification model including a support vector machine classification model in the row direction and a support vector machine classification model in the column direction.

[0044] Specifically, in the embodiment of the present application, a support vector machine (SVM) algorithm is used to determine whether each row and column of pixels in the display panel has defects based on the target defect image. SVM is a two-class classification model, whose basic model is defined as a linear classifier with the largest interval in the feature space, and its learning strategy is to maximize the interval, which can eventually be transformed into a solution to a convex quadratic programming problem. The core of using SVM classification is to find a hyperplane W in the sample space. T x+b=0 is used to distinguish different categories, and then the intra-class distance between the two types of data is calculated. At the same time, the intra-class distance of the hyperplane is optimized to obtain the best classifier through training.

[0045] In one embodiment of the present application, the support vector machine classification model for row direction is defined as: ; Define the support vector machine classification model in the column direction as: ;in; , are row vector samples and column vector samples respectively, , are the normal vectors in the row direction and the normal vector in the column direction, , is the distance between the classification hyperplane and the origin.

[0046] Step S140: According to the row vector samples, column vector samples, the classification labeling results of the row vector samples and the classification labeling results of the column vector samples, respectively obtain the first classification hyperplane of the support vector machine classification model in the row direction and the second classification hyperplane of the support vector machine classification model in the column direction.

[0047] Specifically, the first classification hyperplane and the second classification hyperplane are respectively the optimal hyperplanes of the support vector machine classification model in the row direction and the support vector machine classification model in the column direction. By calculating the optimal solutions in the above models respectively, the optimal hyperplane of the SVM model can be obtained.

[0048] In one embodiment of the present application, when calculating the optimal solution of the above SVM classification model, that is, obtaining the optimal hyperplane, the classification principle is first defined. For the SVM classification model in the row direction, the definition is It means Ri_y =1, Indicates Ri_y = -1; the above classification principle can be transformed into Indicates that the classification is correct. Indicates classification error. For the column-wise SVM classification model, define means Cj_y = 1, Indicates Cj_y = -1; the above classification principle can be transformed into Indicates that the classification is correct. Indicates a classification error.

[0049] According to the above classification principle, mathematical models of hard margin support vector machines in row direction and column direction are established respectively, wherein the mathematical model of hard margin support vector machine in row direction is: ; The mathematical model of the column-wise hard-margin support vector machine is: ; The mathematical model of the hard-margin support vector machine in the row direction and the column direction is solved to obtain the optimal parameters of the first classification hyperplane , and the optimal parameters of the second classification hyperplane .

[0050] Specifically, taking the row-wise SVM classification model as an example, the mathematical model of the row-wise hard-margin support vector machine describes how to find the optimal hyperplane to distinguish the two categories in the data set. The goal of this optimization problem is to maximize the interval between the two categories of data while ensuring that all data points are correctly classified. The above mathematical model includes two parts: the optimization objective and the constraint conditions. The optimization objective indicates that the weight vector W needs to be found. r and bias b r , so that for all samples x i , W r T x i +b r With W r Norm of ||W r The ratio of || is large, which represents the interval between the two types of data. r The norm of is defined as ; here || W r || represents the weight vector W r The Euclidean norm of r The constraint ensures that for all samples x i All are correctly classified.

[0051] During the calculation process, the above optimization objectives are scaled and defined: ; The Lagrangian function is defined as: ; Using the constraints and the gradient being zero, we can get the following solution: ; ;in, is a support vector, is the classification labeling result of the corresponding support vector.

[0052] By adjusting W r and b r And choose the appropriate Lagrangian factor , we can get the optimal parameters of the first classification hyperplane Similarly, the optimal parameters of the second classification hyperplane are The calculation process of refers to the calculation process of the SVM classification model in the row direction, which will not be described in detail in the embodiments of the present application.

[0053] Step S150: Generate a defect detection model storing the first classification hyperplane parameters and the second classification hyperplane parameters.

[0054] Specifically, the optimal parameters calculated above are and The parameters are saved as the first classification hyperplane parameters and the second classification hyperplane parameters respectively, and a classification model for defect detection is generated. When the trained model parameters are used to detect defects on the display panel, for each row and each column in the display image of the display panel, the corresponding SVM model is used to determine whether there is a defect, and the overlapped part of the defective row and defective column is the final detected defect position.

[0055] The method for constructing a defect detection model provided by the embodiment of the present application is to obtain a target defect image of a display panel; construct a row vector sample and a column vector sample according to the brightness data of each pixel in the target defect image, and classify and annotate the row vector sample and the column vector sample respectively; construct a classification model for defect detection, the classification model includes a support vector machine classification model in the row direction and a support vector machine classification model in the column direction; obtain the first classification hyperplane of the support vector machine classification model in the row direction and the second classification hyperplane of the support vector machine classification model in the column direction according to the row vector sample, the column vector sample, the classification and annotation results of the row vector sample and the classification and annotation results of the column vector sample respectively; generate a defect detection model storing the first classification hyperplane parameters and the second classification hyperplane parameters. In this way, a detection model that can automatically detect defect features can be generated to classify the brightness data of pixels in the defect image and realize the detection of local defects in the display panel.

[0056] See also Figure 4 , is a defect detection method provided in an embodiment of the present application, the method uses a defect detection model constructed by the method described in the above embodiment, and the defect detection method specifically includes the following steps: Step S210: obtaining an image to be inspected of the display panel to be inspected.

[0057] Step S220: inputting the image to be inspected into the defect detection model to perform defect detection on the display panel to be inspected.

[0058] Specifically, the brightness data of each row of pixels in the image to be detected is a new row vector sample XN_Ri, and the brightness data of each column of pixels is a new column vector sample XN_Ci; when using the trained SVM classification model for detection, W r T XN_Ri+b r >0 means the row is a defective row, W r T XN_Ri+b r <0 means the row is a non-defective row, W c T XN_Ci+b c >0 means the column is a defective column, W c TXN_Ci+b c <0 means the column is a non-defective column.

[0059] Step S230: when it is detected that the display panel to be inspected has defects, the detected defective rows and defective columns are output.

[0060] In one embodiment, the defect detection method further includes: acquiring the position of the defect area according to the output defect row and defect column.

[0061] Specifically, since the overlapping part of the defective row and the defective column is the defective area finally detected, the four coordinates of the upper left, lower left, upper right and lower right of the defective area can be generated according to the number of defective rows and defective columns in the defective area to locate the defective area.

[0062] The defect detection method provided in the embodiment of the present application obtains the image to be detected of the display panel to be detected; inputs the image to be detected into the defect detection model to perform defect detection on the display panel to be detected; when the display panel to be detected is detected to have defects, the detected defect rows and defect columns are output. In this way, it is possible to quickly detect whether there is a defective area on the display panel, locate the defective area, and thus improve the efficiency of mura compensation, improve the uniformity of screen brightness, and improve the yield of display products.

[0063] Based on the same inventive concept as the aforementioned embodiments, an embodiment of the present application further provides a defect detection device, including a processor and a memory storing a computer program, and when the processor runs the computer program, the steps of the method described in the aforementioned embodiments are implemented.

[0064] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] In this document, the terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of elements other than those listed and may also include additional elements not expressly listed.

[0066] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for constructing a defect detection model, characterized in that: include: Acquire a target defect image of a display panel; Constructing row vector samples and column vector samples according to the brightness data of each pixel in the target defect image, and classifying and labeling the row vector samples and column vector samples respectively; Constructing a classification model for defect detection, wherein the classification model includes a support vector machine classification model in a row direction and a support vector machine classification model in a column direction; According to the row vector samples, the column vector samples, the classification labeling results of the row vector samples and the classification labeling results of the column vector samples, respectively obtaining a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction; A defect detection model storing the first classification hyperplane parameters and the second classification hyperplane parameters is generated.

2. The method for constructing a defect detection model according to claim 1, characterized in that: The step of acquiring a target defect image of a display panel comprises: Acquiring an original defect image of the display panel; Determine the area with defects in the original defect image as the ROI area; The image of the ROI area is preprocessed to obtain the target defect image.

3. The method for constructing a defect detection model according to claim 2, characterized in that: Preprocessing the image of the ROI region includes: adjusting the color scale of the ROI region to enhance the contrast of the target defect image; The step of adjusting the color scale of the image in the ROI area includes: Define calculation parameters in the color scale adjustment, the calculation parameters include: the black field threshold Sin and the white field threshold Hin in the ROI area, the Gamma adjustment factor Mt required for adjustment, and the adjusted output black field threshold Sout and the output white field threshold Hout; The scaling factor V1 is calculated according to the calculation parameters, and the adjustment coefficient V2 after Gamma adjustment is calculated according to the scaling factor V1, wherein ; ; The Pin is the original brightness data of the pixels in the ROI area; An output mapping function is established according to the adjustment coefficient V2 to obtain the output brightness data of the pixels in the ROI area after the color scale adjustment. The output mapping function is: 。 4. The method for constructing a defect detection model according to claim 1, characterized in that: The step of constructing row vector samples and column vector samples according to the brightness data of each pixel in the target defect image includes: The brightness data of each row of pixels in the target defect image is taken as a row vector sample, and the row vector sample set X_Ri = [pi1, pi2, ..., piM], M = COL, i = 1, ..., ROW; The brightness data of each column of pixels in the target defect image is taken as a column vector sample, and the column vector sample set X_Cj = [p1j, p2j, ..., pNj], N = ROW, j = 1, ..., COL; Among them, ROW is the number of rows of pixels in the target defect image, COL is the number of columns of pixels in the target defect image, and piM and pNj are the brightness of pixels in the target defect image.

5. The method for constructing a defect detection model according to claim 1, characterized in that: The classifying and labeling the row vector samples and the column vector samples respectively includes: Determining whether the pixel is a defective pixel according to the brightness data of the pixel in the row vector sample and the column vector sample; Classifying and labeling the row vector samples and the column vector samples according to the determination result of the defective pixels; Among them, when the classification labeling result of the row vector sample Ri_y = 1, it means that the i-th row in the target defect image is a defective row, and when Ri_y = -1, it means that the i-th row is a non-defective row; when the classification standard result of the column vector sample Cj_y = 1, it means that the j-th column in the target defect image is a defective column, and when Cj_y = -1, it means that the j-th column is a non-defective column.

6. The method for constructing a defect detection model according to claim 5, characterized in that: The support vector machine classification model in the row direction is: ; The support vector machine classification model in the column direction is: ; in; , are row vector samples and column vector samples respectively, , are the normal vectors in the row direction and the normal vector in the column direction, , is the distance between the classification hyperplane and the origin.

7. The method for constructing a defect detection model according to claim 6, characterized in that: Obtaining a first classification hyperplane of the support vector machine classification model in the row direction and a second classification hyperplane of the support vector machine classification model in the column direction, comprising: The mathematical models of hard-margin support vector machines in row and column directions are established respectively. The mathematical model of the hard-margin support vector machine in the row direction is: ; The mathematical model of the column-wise hard-margin support vector machine is: ; Solve the mathematical model of the hard margin support vector machine in the row direction and column direction respectively to obtain the optimal parameters of the first classification hyperplane , and the optimal parameters of the second classification hyperplane .

8. A defect detection method, characterized in that: A defect detection model constructed using the method according to any one of claims 1 to 8, the method comprising: Acquire an image to be detected of a display panel to be detected; Inputting the image to be inspected into the defect detection model to perform defect detection on the display panel to be inspected; When it is detected that the display panel to be inspected has defects, the detected defective rows and defective columns are output.

9. The defect detection method according to claim 8, characterized in that: The method further comprises: Get the location of the defect area based on the output defect row and defect column.

10. A defect detection device, characterized in that: The method comprises a processor and a memory storing a computer program, and when the processor runs the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images